Analog Tunnel Memory Based on Programmable Metallization for Passive Neuromorphic Circuits

Abstract

Although experimental implementations of memristive crossbar arrays have indicated the potential of these networks for in-memory computing, their performance is generally limited by an intrinsic variability on the device level as a result of the stochastic formation of conducting filaments. A tunnel-type memristive device typically exhibits small switching variations, owing to the relatively uniform interface effect. However, the low mobility of oxygen ions and large depolarization field result in slow operation speed and poor retention. Here, we demonstrate a quantum-tunneling memory with Ag-doped percolating systems, which possesses desired characteristics for large-scale artificial neural networks. The percolating layer suppresses the random formation of conductive filaments, and the nonvolatile modulation of the Fowler−Nordheim tunneling current is enabled by the collective movement of active Ag nanocrystals with high mobility and a minimal depolarization field. Such devices simultaneously possess electroforming-free characteristics, record low switching variabilities (temporal and spatial variation down to 1.6 and 2.1%, respectively), nanosecond operation speed, and long data retention (>10 4 s at 85 °C). Simulations prove that passive arrays with our analog memory of large current−voltage nonlinearity achieve a high write and recognition accuracy. Thus, our discovery of the unique tunnel memory contributes to an important step toward realizing neuromorphic circuits.

Criticality and Neuromorphic Sensing in a Single Memristor

“…These results suggest that the filament dynamics in memristors can be tuned into critical regimes, which may be utilized to optimize signal detection in sensory systems. We also note that the work presented here is complementary to the results obtained on memristor-based electronic synapses, , artificial neurons, , and other neuromorphic devices, and together those will contribute to low-cost bioplausible neuromorphic systems.…”

Self‐Rectifying Memristors for Three‐Dimensional In‐Memory Computing

“…[136] Researchers typically simulate the long-term potentiation (LTP) or long-term depression (LTD) behavior of biological synapses by applying continuous positive or negative voltage pulses to the device to continuously increase or decrease its conductance state of the device. [62][63][64]93,[137][138][139][140][141]164] Cheong et al proposed a neuromodulation-inspired stashing system for spiking neural networks using a 32 × 32 SRM array (Figure 10a). [64] Based on the specific energy band structure design, the trilayer (HfO 2 /Nb 2 O 5 /HfO 2 )-based SRM array (Figure 10b) exhibits desirable analog switching characteristics of stable 4-bit (16 levels) conductance between 0.45 and 7.45 nS (Figure 10c-e).…”

“…[ 136 ] Researchers typically simulate the long‐term potentiation (LTP) or long‐term depression (LTD) behavior of biological synapses by applying continuous positive or negative voltage pulses to the device to continuously increase or decrease its conductance state of the device. [ 62–64,93,137–141,164 ]…”